Module 02360 (2004)

Syllabus page 2004/2005

06-02360
Introduction to Neural Networks

Level 2/I

John Bullinaria
10 credits in Semester 1

Links | Outline | Aims | Outcomes | Prerequisites | Teaching | Assessment | Books | Detailed Syllabus


The Module Description is a strict subset of this Syllabus Page. (The University module description has not yet been checked against the School's.)

Relevant Links

See Neural Networks Web-page for module material and further useful links.


Outline

This module provides an introduction to basic neurobiology, discusses the main neural network architectures and learning algorithms, and presents a number of neural network applications. Particular models covered include McCulloch Pitts Neurons, Single Layer Perceptrons, Multi-Layer Perceptrons, Radial Basis Function Networks, Committee Machines, Kohonen Self-Organising Maps, and Learning Vector Quantization.


Aims

The aims of this module are to:

  • introduce the main fundamental principles and techniques of neural network systems
  • investigate the principal neural network models and applications

Learning Outcomes

On successful completion of this module, the student should be able to: Assessed by:
1describe the relation between real brains and simple artificial neural network models Examination
2explain and contrast the most common architectures and learning algorithms for Multi-Layer Perceptrons, Radial-Basis Function Networks, Committee Machines, and Kohonen Self-Organising Maps Examination
3discuss the main factors involved in achieving good learning and generalization performance in neural network systems Examination, assignment
4identify the main implementational issues for common neural network systems Examination, assignment
5evaluate the practical considerations in applying neural networks to real classification and regression problems Examination, assignment

Restrictions, Prerequisites and Corequisites

Restrictions:

None

Prerequisites:

None

Co-requisites:

None


Teaching

Teaching Methods:

2 hrs of lectures per week plus labs

Contact Hours:

24


Assessment

  • Supplementary (where allowed): As the sessional assessment
  • 2 hour examination (70%), continuous assessment (30%). Resit by written examination only with the continuous assessment mark carried forward.

Recommended Books

TitleAuthor(s)Publisher, Date
An Introduction to Neural NetworksK GurneyRoutledge, 1997
Neural Networks: A Comprehensive FoundationS HaykinPrentice Hall, 1999
Neural Networks for Pattern RecognitionC M BishopOxford University Press, 1995
The Essence of Neural NetworksR CallanPrentice Hall Europe, 1999
Introduction to Neural NetworksR Beale & T JacksonIOP Publishing, 1990
An Introduction to the Theory of Neural ComputationJ Hertz, A Krogh & R G PalmerAddison Wesley, 1991
Principles of Neurocomputing for Science and EngineeringF M Ham & I KostanicMcGraw Hill, 2001

Detailed Syllabus

  1. Introduction to Neural Networks and their History.
  2. Biological Neurons and Neural Networks. Artificial Neurons.
  3. Networks of Artificial Neurons. Single Layer Perceptrons.
  4. Learning and Generalization in Single Layer Perceptrons.
  5. Hebbian Learning. Gradient Descent Learning.
  6. The Generalized Delta Rule. Practical Considerations.
  7. Learning in Multi-Layer Perceptrons. Back-Propagation.
  8. Learning with Momentum. Conjugate Gradient Learning.
  9. Bias and Variance. Under-Fitting and Over-Fitting.
  10. Improving Generalization.
  11. Applications of Multi-Layer Perceptrons.
  12. Radial Basis Function Networks: Introduction.
  13. Radial Basis Function Networks: Algorithms.
  14. Radial Basis Function Networks: Applications.
  15. Committee Machines.
  16. Self Organizing Maps: Fundamentals.
  17. Self Organizing Maps: Algorithms and Applications.
  18. Learning Vector Quantisation.
  19. Overview of More Advanced Topics.

Last updated: 8 December 2003

Source file: /internal/modules/COMSCI/2004/xml/02360.xml

Links | Outline | Aims | Outcomes | Prerequisites | Teaching | Assessment | Books | Detailed Syllabus